When a high-profile psychologist reviewed her newly published paper in PLOS ONE, she was dismayed to notice multiple formatting errors.
So she contacted the journal to find out what had gone wrong, especially since checking the page proofs would have spotted the problem immediately. The authors were surprised to learn that it was against the journal’s policy to provide authors page proofs. Could this partly explain PLOS ONE’s high rate of corrections?
Issuing frequent corrections isn’t necessarily a bad thing, since it can indicate that the journal is responsive to fixing published articles. But the rate of corrections at PLOS ONE is notably high.
According to an analysis published yesterday on a blog by Mark Dingemanse, a researcher at the Max Planck Institute for Psycholinguistics, between May and June, 2015, the journal published 8466 articles and 474 corrections. Of course, many of those corrections were issued for articles published before that specific time period, but that still works out to a correction rate of more than 5% of the journal’s total output.
In contrast, Medline data show that more than 800,000 papers were published in the 2015 fiscal year, as well as more than 12,000 errata (again, not all errata were for papers published in 2015); this gives a correction rate of around 1.5%.
— Dorothy Bishop (@deevybee) August 4, 2016
Despite any past republication instances and regardless of fault, current PLOS policy does not permit republication unless confidential or copyrighted information has been published and needs to be purged from the publication record. At this time, we do not provide proofs during the production process of our manuscripts.
It went further to say:
…we do not publish corrections for errors that do not affect the scientific content of the article. Because these are not content errors, there is nothing further we can do.
When queried about the policy, Bishop told us:
This is very unusual and out of line with normal publishing practice — and, if my twitter feed is anything to go by, it has led to numerous problems. One person even noted that PLOS One had misspelt their name on a paper and then said it could not be corrected.
After hearing from the journal, Bishop posted this comment under her paper:
I regret that the formatting of the pdf version of this article impairs readability.
The authors had formatted the article so that the 27 statements were distinctive, but we were asked to remove that formatting from the manuscript version of the document.
At this point we did not realise that PLOS One no longer provides authors with page proofs, since this is a change in their previous practice. Had we been able to check proofs, we would have pointed out that in the pdf version, the statements do not stand out and are in some cases run in with the prior heading. The “Supplementary Comments” on each statement have more prominence than the statements themselves.
PLOS One editorial staff tell us they are not able to correct formatting errors.
This isn’t the first PLOS ONE paper with an error stemming from typesetting problems. As Dingemanse writes:
In the period May-July 2015, PLOS ONE published a total of 474 corrections. Over a quarter of these (132) indicate that the error was introduced at the typesetting stage, i.e. beyond the control of the authors.
According to Dingemanse, the numbers for the whole of 2015 are even worse: The journal published 30,790 research papers and issued 1939 corrections in 2015 (again, not necessarily for papers published in 2015), which bring its correction rate to just under 6.3%. Additionally, 415 (21.2%) of those corrections acknowledge publisher errors.
Furthermore, Dingemanse says, the rate of corrections due to publisher errors seem to be rising: 13% of corrections published in 2014 cite a publisher error, and the number is as low as 4% for corrections in 2013.
Statistics released by the US National Library of Medicine suggest that the number of errata increased by 29% — from 9,602 to 12,344 — between the 2014 and 2015 fiscal years.
However, a 2013 study by Daniele Fanelli, a researcher at the Meta-Research Innovation Center at Stanford University in California, noted that the proportion of published errata has not increased for decades:
…despite a steady increase in the number of publications covered by the [Web of Science], the proportion of errata has remained relatively constant since the 1970s (and arguably since the 1950s).
In response to Bishop, other Twitter users also posted their experiences:
in my case they actually incorrectly photoshopped axis labels back into my figure in the correction
— Michael C. Frank (@mcxfrank) August 4, 2016
Others have questioned whether PLOS ONE’s massive editorial challenge — publishing more than 30,000 articles — raises the chances that mistakes will happen. (Remember the paper that cited “the Creator?”) As David Crotty wrote in March in Scholarly Kitchen:
…what about editorial decision-making? PLOS ONE has some 6,100 editors. Rather than funneling everything through an Editor in Chief, the peer review and decision-making process is spread broadly…There is no consistent level of quality control because there are 6,100 different sets of standards being used and no central point where they come together.
PLOS ONE were unable to comment on the situation by press time, but a spokesperson from the organization told us to
Look for a more thorough, accurate, meaningful and in context blog post from PLOS in the next week or so.
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.